void set_sample_label (DATASET *dset)
{
GtkWidget *dlabel;
char tmp[256];
int tsubset;
if (mdata == NULL) {
return;
}
/* set the sensitivity of various menu items */
time_series_menu_state(dataset_is_time_series(dset));
panel_menu_state(dataset_is_panel(dset));
ts_or_panel_menu_state(dataset_is_time_series(dset) ||
dataset_is_panel(dset));
flip(mdata->ui, "/menubar/Data/DataTranspose", !dataset_is_panel(dset));
tsubset = dset->t1 > 0 || dset->t2 < dset->n - 1;
/* construct label showing summary of dataset/sample info
(this goes at the foot of the window)
*/
if (complex_subsampled() && !tsubset && dataset_is_cross_section(dset)) {
sprintf(tmp, _("Undated: Full range n = %d; current sample"
" n = %d"), get_full_length_n(), dataset->n);
gtk_label_set_text(GTK_LABEL(mdata->status), tmp);
} else if (complex_subsampled() && dataset_is_panel(dset)) {
char t1str[OBSLEN], t2str[OBSLEN];
const char *pdstr = get_pd_string(dset);
ntodate(t1str, dset->t1, dset);
ntodate(t2str, dset->t2, dset);
sprintf(tmp, _("%s; sample %s - %s"), _(pdstr), t1str, t2str);
gtk_label_set_text(GTK_LABEL(mdata->status), tmp);
} else {
char t1str[OBSLEN], t2str[OBSLEN];
const char *pdstr = get_pd_string(dset);
if (tsubset && calendar_data(dset)) {
/* it's too verbose to print both full range and sample */
ntodate(t1str, dset->t1, dset);
ntodate(t2str, dset->t2, dset);
sprintf(tmp, _("%s; sample %s - %s"), _(pdstr), t1str, t2str);
gtk_label_set_text(GTK_LABEL(mdata->status), tmp);
} else {
ntodate(t1str, 0, dset);
ntodate(t2str, dset->n - 1, dset);
sprintf(tmp, _("%s: Full range %s - %s"), _(pdstr),
t1str, t2str);
if (tsubset) {
gchar *fulltext;
ntodate(t1str, dset->t1, dset);
ntodate(t2str, dset->t2, dset);
fulltext = g_strdup_printf(_("%s; sample %s - %s"), tmp, t1str, t2str);
gtk_label_set_text(GTK_LABEL(mdata->status), fulltext);
g_free(fulltext);
} else {
gtk_label_set_text(GTK_LABEL(mdata->status), tmp);
}
}
}
/* construct label with datafile name (this goes above the
data series window) */
dlabel = g_object_get_data(G_OBJECT(mdata->main), "dlabel");
if (dlabel != NULL) {
if (strlen(datafile) > 2) {
/* data file open already */
const char *p = strrchr(datafile, SLASH);
gchar *trfname;
if (p != NULL) {
trfname = my_filename_to_utf8(p + 1);
} else {
trfname = my_filename_to_utf8(datafile);
}
strcpy(tmp, " ");
if (data_status & SESSION_DATA) {
sprintf(tmp + 1, "Imported %s", trfname);
} else if (data_status & MODIFIED_DATA) {
sprintf(tmp + 1, "%s *", trfname);
} else {
sprintf(tmp + 1, "%s", trfname);
}
gtk_label_set_text(GTK_LABEL(dlabel), tmp);
g_free(trfname);
} else if (data_status & MODIFIED_DATA) {
strcpy(tmp, _(" Unsaved data "));
gtk_label_set_text(GTK_LABEL(dlabel), tmp);
}
}
if (complex_subsampled() || dset->t1 > 0 ||
dset->t2 < dset->n - 1) {
restore_sample_state(TRUE);
}
console_record_sample(dataset);
}
GtkWidget *build_var_popup (void)
{
const char *items[] = {
N_("Display values"),
N_("Summary statistics"),
N_("Time series plot"),
N_("Panel plot..."),
N_("Frequency distribution"),
N_("Boxplot"),
N_("Correlogram"),
N_("Periodogram"),
N_("Edit attributes"),
N_("Edit values"),
N_("Copy to clipboard"),
N_("Delete"),
NULL,
N_("Add log"),
N_("Add difference"),
N_("Define new variable...")
};
GtkWidget *menu;
GtkWidget *item;
int i, n = G_N_ELEMENTS(items);
int real_panel = multi_unit_panel_sample(dataset);
menu = gtk_menu_new();
for (i=0; i<n; i++) {
if (items[i] == NULL) {
item = gtk_separator_menu_item_new();
gtk_widget_show(item);
gtk_menu_shell_append(GTK_MENU_SHELL(menu), item);
continue;
}
if (real_panel && (i == 2 || i == 5)) {
/* don't offer regular ts or boxplot */
continue;
}
if (!real_panel && i == 3) {
/* don't offer panel plot */
continue;
}
if ((i == 6 || i == 7) && !dataset_is_time_series(dataset)) {
continue;
}
if ((i == 2 || i == 14) && !extended_ts(dataset)) {
continue;
}
if (i == 5 && dataset_is_time_series(dataset)) {
/* skip boxplot option */
continue;
}
item = gtk_menu_item_new_with_label(_(items[i]));
g_signal_connect(G_OBJECT(item), "activate",
G_CALLBACK(var_popup_click),
_(items[i]));
gtk_widget_show(item);
gtk_menu_shell_append(GTK_MENU_SHELL(menu), item);
}
return menu;
}
DATASET *midas_aux_dataset (const int *list,
const DATASET *dset,
int *err)
{
DATASET *mset = NULL;
gretlopt opt = 0;
int mpd, pd = dset->pd;
int T, m = list[0];
int yr, mon;
int daily = 0;
if (m < 3 || gretl_list_has_separator(list)) {
*err = E_INVARG;
} else if (!dataset_is_time_series(dset)) {
*err = E_INVARG;
} else if (pd != 1 && pd != 4 && pd != 12) {
/* host dataset should be annual, quarterly or monthly */
*err = E_PDWRONG;
}
if (*err) {
return NULL;
}
if (pd == 1) {
/* annual: midas series should be quarterly or monthly */
if (m != 4 && m != 12) {
*err = E_INVARG;
} else {
mpd = m;
}
} else if (pd == 4) {
/* quarterly: midas series should be monthly or daily */
if (m == 3) {
mpd = 12;
} else if (m == midas_days_per_period(5, 4)) {
mpd = 5;
} else if (m == midas_days_per_period(6, 4)) {
mpd = 6;
} else if (m == midas_days_per_period(7, 4)) {
mpd = 7;
} else {
*err = E_INVARG;
}
} else {
/* monthly: midas series should be daily */
if (m == midas_days_per_period(5, 12)) {
mpd = 5;
} else if (m == midas_days_per_period(6, 12)) {
mpd = 6;
} else if (m == midas_days_per_period(7, 12)) {
mpd = 7;
} else {
*err = E_INVARG;
}
}
if (*err) {
return NULL;
}
if (!gretl_is_midas_list(list, dset)) {
gretl_warnmsg_set("The argument does not seem to be a MIDAS list");
}
T = sample_size(dset) * m;
if (mpd >= 5 && mpd <= 7) {
/* we'll add markers for daily dates */
daily = 1;
opt = OPT_M;
}
mset = create_auxiliary_dataset(1, T, opt);
if (mset == NULL) {
*err = E_ALLOC;
}
if (!*err) {
char *p, obs[OBSLEN];
int nonex, qtr = 0;
int i, t, s, m3 = 0;
mset->pd = mpd;
mset->structure = TIME_SERIES;
strcpy(mset->varname[0], dset->varname[list[1]]);
p = strrchr(mset->varname[0], '_');
if (p != NULL) *p = '\0';
ntodate(obs, dset->t1, dset);
if (mpd == 4) {
sprintf(mset->stobs, "%d:1", atoi(obs));
} else if (mpd == 12) {
sprintf(mset->stobs, "%d:01", atoi(obs));
}
if (daily && pd == 4) {
m3 = m / 3;
}
/* loop across observations in low-frequency dataset */
s = 0;
for (t=dset->t1; t<=dset->t2; t++) {
if (daily) {
ntodate(obs, t, dset);
sscanf(obs, "%d:%d", &yr, &mon);
if (pd == 4) {
qtr = mon;
}
}
/* read data right-to-left! */
for (i=m; i>0; i--) {
int vi = list[i];
if (daily) {
if (pd == 4) {
mon = quarter_to_month(qtr, m, m-i+1);
nonex = daily_index_to_date(mset->S[s], yr, mon,
(m-i) % m3, mpd);
} else {
nonex = daily_index_to_date(mset->S[s], yr, mon,
m-i, mpd);
}
if (nonex) {
/* skip any non-existent daily dates */
mset->t2 -= 1;
} else {
mset->Z[0][s++] = dset->Z[vi][t];
}
} else {
mset->Z[0][s++] = dset->Z[vi][t];
}
}
}
if (daily) {
strcpy(mset->stobs, mset->S[0]);
strcpy(mset->endobs, mset->S[mset->t2]);
mset->markers = DAILY_DATE_STRINGS;
}
mset->sd0 = get_date_x(mset->pd, mset->stobs);
if (!daily) {
ntodate(mset->endobs, mset->t2, mset);
}
}
return mset;
}
int stata_export (const char *fname,
const int *list,
gretlopt opt,
const DATASET *dset)
{
dta_hdr hdr;
char buf[96];
char timevar[16];
ssize_t w = 0;
guint8 *types;
double xit;
gint32 i32, t32 = 0;
gint16 i16;
guint8 i8;
int missing;
int add_time = 0;
int i, j, t, fd;
int strvars = 0;
int nv = 0;
int err = 0;
fd = gretl_open(fname, O_WRONLY | O_CREAT | O_TRUNC, 0644);
if (fd < 0) {
return E_FOPEN;
}
types = make_types_array(dset, list, &nv);
if (types == NULL) {
close(fd);
return E_ALLOC;
}
*timevar = '\0';
if (dataset_is_time_series(dset)) {
t32 = get_stata_t0(dset, timevar, &add_time);
}
dta_hdr_init(&hdr, dset, nv + (add_time > 0));
dta_hdr_write(&hdr, fd, &w);
/*
typlist nvar byte array
varlist 33*nvar char array
srtlist 2*(nvar+1) int array
fmtlist 12*nvar char array
lbllist 33*nvar char array
*/
/* typlist */
if (add_time) {
i8 = STATA_LONG;
w += write(fd, &i8, 1);
}
for (j=0; j<nv; j++) {
w += write(fd, &types[j], 1);
}
/* varlist (names) */
if (add_time) {
memset(buf, 0, 33);
strcat(buf, timevar);
w += write(fd, buf, 33);
}
for (i=1; i<dset->v; i++) {
if (include_var(list, i)) {
memset(buf, 0, 33);
strcat(buf, dset->varname[i]);
w += write(fd, buf, 33);
}
}
/* srtlist */
i16 = 0;
if (add_time) {
w += write(fd, &i16, 2);
}
for (j=0; j<=nv; j++) {
w += write(fd, &i16, 2);
}
/* fmtlist */
if (add_time) {
memset(buf, 0, 12);
sprintf(buf, "%%t%c", *timevar);
w += write(fd, buf, 12);
}
for (j=0; j<nv; j++) {
memset(buf, 0, 12);
if (types[j] == STATA_BYTE || types[j] == STATA_INT) {
strcpy(buf, "%8.0g");
} else if (types[j] == STATA_LONG) {
strcpy(buf, "%12.0g");
} else {
strcpy(buf, "%9.0g");
}
w += write(fd, buf, 12);
}
/* lbllist */
if (add_time) {
memset(buf, 0, 33);
w += write(fd, buf, 33);
}
for (i=1; i<dset->v; i++) {
if (include_var(list, i)) {
memset(buf, 0, 33);
if (is_string_valued(dset, i)) {
sprintf(buf, "S%d", i);
strvars++;
}
w += write(fd, buf, 33);
}
}
/* Variable labels */
if (add_time) {
memset(buf, 0, 81);
make_timevar_label(buf, timevar);
w += write(fd, buf, 81);
}
for (i=1; i<dset->v; i++) {
if (include_var(list, i)) {
memset(buf, 0, 81);
asciify_to_length(buf, series_get_label(dset, i), 80);
w += write(fd, buf, 81);
}
}
/* "Expansion fields" (a Stata-private mystery) */
i8 = 0;
for (i=0; i<5; i++) {
w += write(fd, &i8, 1);
}
/* Data */
for (t=dset->t1; t<=dset->t2; t++) {
if (add_time == TIME_AUTO) {
w += write(fd, &t32, 4);
t32++;
} else if (add_time == TIME_CAL) {
t32 = stata_date(t, dset);
w += write(fd, &t32, 4);
}
j = 0;
for (i=1; i<dset->v; i++) {
if (include_var(list, i)) {
xit = dset->Z[i][t];
missing = xna(xit);
if (types[j] == STATA_BYTE) {
i8 = missing ? STATA_BYTE_NA : xit;
w += write(fd, &i8, 1);
} else if (types[j] == STATA_INT) {
i16 = missing ? STATA_INT_NA : xit;
w += write(fd, &i16, 2);
} else if (types[j] == STATA_LONG) {
i32 = missing ? STATA_LONG_NA : xit;
w += write(fd, &i32, 4);
} else {
xit = missing ? STATA_DOUBLE_NA : xit;
w += write(fd, &xit, 8);
}
j++;
}
}
}
/* Value labels */
if (strvars > 0) {
for (i=1; i<dset->v; i++) {
if (include_var(list, i) &&
is_string_valued(dset, i)) {
dta_value_labels_write(fd, dset, i, &w);
}
}
}
close(fd);
free(types);
fprintf(stderr, "stata_export: wrote %d bytes\n", (int) w);
return err;
}
int gnumeric_get_data (const char *fname, int *list, char *sheetname,
DATASET *dset, gretlopt opt, PRN *prn)
{
int gui = (opt & OPT_G);
wbook gbook;
wbook *book = &gbook;
wsheet gsheet;
wsheet *sheet = &gsheet;
int sheetnum = -1;
int obscol = 0;
DATASET *newset;
int err = 0;
newset = datainfo_new();
if (newset == NULL) {
pputs(prn, _("Out of memory\n"));
return 1;
}
wsheet_init(sheet);
gretl_push_c_numeric_locale();
if (wbook_get_info(fname, list, sheetname, book, prn)) {
pputs(prn, _("Failed to get workbook info"));
err = 1;
goto getout;
}
wbook_print_info(book);
if (book->nsheets == 0) {
pputs(prn, _("No worksheets found"));
err = 1;
goto getout;
}
if (gui) {
if (book->nsheets > 1) {
wsheet_menu(book, 1);
sheetnum = book->selected;
} else {
wsheet_menu(book, 0);
sheetnum = 0;
}
} else {
err = wbook_check_params(book);
if (err) {
gretl_errmsg_set(_("Invalid argument for worksheet import"));
} else if (book->selected >= 0) {
sheetnum = book->selected;
} else {
sheetnum = 0;
}
}
if (book->selected == -1) {
/* canceled */
err = -1;
}
if (!err && sheetnum >= 0) {
fprintf(stderr, "Getting data...\n");
err = wsheet_setup(sheet, book, sheetnum);
if (!err) {
err = wsheet_get_data(fname, sheet, &obscol, prn);
if (err) {
fprintf(stderr, "wsheet_get_data returned %d\n", err);
} else {
wsheet_print_info(sheet);
book->flags |= sheet->flags;
}
}
}
if (err) {
goto getout;
} else {
int r0 = 1; /* the first data row */
int i, j, t;
int ts_markers = 0;
int merge = (dset->Z != NULL);
char **ts_S = NULL;
int blank_cols = 0;
int missvals = 0;
int pd = 0;
if (obscol) {
book_set_obs_labels(book);
if (sheet->text_cols == 0) {
sheet->text_cols = 1;
}
} else if (sheet->text_cols > 0) {
/* string-valued variable? */
fprintf(stderr, "Problem: sheet->text_cols = %d\n", sheet->text_cols);
}
if (sheet->colheads == 0) {
book_set_auto_varnames(book);
r0 = 0;
}
if (book_numeric_dates(book)) {
fputs("found calendar dates in first imported column\n", stderr);
} else if (obscol) {
fprintf(stderr, "found label strings in first imported column (text_cols = %d)\n",
sheet->text_cols);
} else if (sheet->text_cols > 0) {
fputs("found string-valued variable in first imported column?\n", stderr);
} else {
fputs("check for label strings in first imported column: not found\n", stderr);
}
newset->n = sheet->maxrow - sheet->row_offset;
if (!sheet->colheads) {
pputs(prn, _("it seems there are no variable names\n"));
newset->n += 1;
}
if (book_numeric_dates(book) || obscol) {
pd = importer_dates_check(sheet->label + r0, &book->flags,
newset, prn, &err);
if (pd > 0) {
/* got time-series info from dates/labels */
sheet_time_series_setup(sheet, book, newset, pd);
ts_markers = newset->markers;
ts_S = newset->S;
} else if (!book_numeric_dates(book)) {
if (labels_are_index(sheet->label, newset->n)) {
/* trash the labels */
book_unset_obs_labels(book);
}
}
}
newset->v = sheet->maxcol + 2 - sheet->col_offset - sheet->text_cols;
fprintf(stderr, "newset->v = %d, newset->n = %d\n",
newset->v, newset->n);
/* create import dataset */
err = worksheet_start_dataset(newset);
if (err) {
goto getout;
}
if (book_time_series(book)) {
newset->markers = ts_markers;
newset->S = ts_S;
} else {
dataset_obs_info_default(newset);
}
j = 1;
for (i=1; i<newset->v; i++) {
int s = (sheet->colheads)? 1 : 0;
int k = i - 1 + sheet->text_cols;
double zkt;
if (column_is_blank(sheet, k, newset->n)) {
blank_cols++;
continue;
}
if (sheet->colheads && *sheet->varname[k] != '\0') {
strcpy(newset->varname[j], sheet->varname[k]);
} else {
sprintf(newset->varname[j], "v%d", j);
}
for (t=0; t<newset->n; t++) {
zkt = sheet->Z[k][s++];
if (zkt == -999 || zkt == -9999) {
newset->Z[j][t] = NADBL;
} else {
newset->Z[j][t] = zkt;
}
if (na(newset->Z[j][t])) {
missvals = 1;
}
}
j++;
}
if (blank_cols > 0) {
fprintf(stderr, "Dropping %d apparently blank column(s)\n",
blank_cols);
dataset_drop_last_variables(newset, blank_cols);
}
if (missvals) {
pputs(prn, _("Warning: there were missing values\n"));
}
if (fix_varname_duplicates(newset)) {
pputs(prn, _("warning: some variable names were duplicated\n"));
}
if (book_obs_labels(book) && wsheet_labels_complete(sheet)) {
int offset = (sheet->colheads)? 1 : 0;
dataset_allocate_obs_markers(newset);
if (newset->S != NULL) {
for (t=0; t<newset->n; t++) {
strcpy(newset->S[t], sheet->label[t+offset]);
}
}
}
if (book->flags & BOOK_DATA_REVERSED) {
reverse_data(newset, prn);
}
if (!err && !dataset_is_time_series(newset) && newset->S != NULL) {
/* we didn't time series info above, but it's possible
the observation strings carry such info
*/
import_ts_check(newset);
}
err = merge_or_replace_data(dset, &newset, opt, prn);
if (!err && !merge) {
dataset_add_import_info(dset, fname, GRETL_GNUMERIC);
}
if (!err && gui) {
wbook_record_params(book, list);
}
}
getout:
wbook_free(book);
wsheet_free(sheet);
gretl_pop_c_numeric_locale();
if (err && newset != NULL) {
destroy_dataset(newset);
}
return err;
}
static int leverage_plot (const MODEL *pmod, gretl_matrix *S,
DATASET *dset)
{
FILE *fp;
const double *obs = NULL;
int t, err = 0;
fp = open_plot_input_file(PLOT_LEVERAGE, 0, &err);
if (err) {
return err;
}
if (dataset_is_time_series(dset)) {
obs = gretl_plotx(dset, OPT_NONE);
if (obs == NULL) {
if (fp != NULL) {
fclose(fp);
}
return 1;
}
}
gretl_push_c_numeric_locale();
fputs("set size 1.0,1.0\nset multiplot\nset size 1.0,0.48\n", fp);
fputs("set xzeroaxis\n", fp);
fputs("set nokey\n", fp);
if (obs != NULL) {
leverage_x_range(pmod->t1, pmod->t2, obs, fp);
} else {
fprintf(fp, "set xrange [%g:%g]\n",
pmod->t1 + 0.5, pmod->t2 + 1.5);
}
/* upper plot: leverage factor */
fputs("set origin 0.0,0.50\n", fp);
gnuplot_missval_string(fp);
fputs("set yrange [0:1]\n", fp);
fprintf(fp, "set title '%s'\n", _("leverage"));
fputs("plot \\\n'-' using 1:2 w impulses\n", fp);
for (t=pmod->t1; t<=pmod->t2; t++) {
double h = gretl_matrix_get(S, t - pmod->t1, 0);
if (na(h)) {
if (obs != NULL) {
fprintf(fp, "%g ?\n", obs[t]);
} else {
fprintf(fp, "%d ?\n", t + 1);
}
} else {
if (obs != NULL) {
fprintf(fp, "%g %g\n", obs[t], h);
} else {
fprintf(fp, "%d %g\n", t + 1, h);
}
}
}
fputs("e\n", fp);
/* lower plot: influence factor */
fputs("set origin 0.0,0.0\n", fp);
gnuplot_missval_string(fp);
fputs("set yrange [*:*]\n", fp);
fprintf(fp, "set title '%s'\n", _("influence"));
fputs("plot \\\n'-' using 1:2 w impulses\n", fp);
for (t=pmod->t1; t<=pmod->t2; t++) {
double f = gretl_matrix_get(S, t - pmod->t1, 1);
if (na(f)) {
if (obs != NULL) {
fprintf(fp, "%g ?\n", obs[t]);
} else {
fprintf(fp, "%d ?\n", t + 1);
}
} else {
if (obs != NULL) {
fprintf(fp, "%g %g\n", obs[t], f);
} else {
fprintf(fp, "%d %g\n", t + 1, f);
}
}
}
fputs("e\n", fp);
fputs("unset multiplot\n", fp);
gretl_pop_c_numeric_locale();
return finalize_plot_input_file(fp);
}
